Digital Fyber Focus is our online series hosting top industry players, thought leaders, and experts to share actionable insights and perspectives about different industry topics and react in real-time to current events that are going on in our ecosystem.
The post-IDFA era is here. Apple’s Tracking Transparency (ATT) has been enforced and without the IDFA taken for granted, advertisers will need to rely on alternative sustainable solutions. As targeting shifts from its granular-level, and true personalization of ads is not as accessible, contextual targeting has emerged as a viable solution.
With over two million apps on the App Store, games make up 21% of active apps. Competition is fierce. With that said, discovery is key in mobile marketing and guiding users to successfully downloading your app can dramatically influence your user acquisition and monetization strategy. But how does contextual targeting play a part in this?
In this Digital Fyber Focus session, Itai Cohen, VP Marketing & Corporate Strategy, and Joseph Iris, RTB Product Lead at Persona.ly discuss how contextual targeting has demonstrated itself as an alternative solution, enabling privacy-aware targeting. Fyber’s contextual targeting solutions focus on user engagement, app-level, and game mode signals. Persona.ly, on the other hand, have used contextual targeting as a means to target users across games through app store optimization and achieve user acquisition in this pool of similar users.
Here are our three key takeaways from this conversation:
App store optimization in a contextual environment
Apps optimize their pages on the app store by using specific keywords and creatives that will help catch a user’s attention. By utilizing the content space publishers have available in the app store, namely the themes within the game, the features and its benefits, you can increase the likelihood of acquiring users – because apps that are similar, have similar users.
To do this, Persona.ly use natural language processing (NLP) techniques to manipulate text into representing users or apps. The algorithm then measures the distance between apps by using vectors to create a reliable representation of the relationships and proximity between two apps.
By identifying games that carry similar themes in the app store descriptions, you can most likely buy users who would like the game you’re advertising. For example, if you’re trying to promote a match 3 game, identifying games that are similar through app store descriptions, increases the chance of targeting and finding similar users who would download a similar game from the puzzle genre.
Pro tip: The key to remaining relevant is to keep updating your texts. You can think of it as SEO for your app. These three or four sentences of text also act as your elevator pitch to your game and capturing a users’ attention in this moment is critical to achieving installs.
Measuring contextual distance using neural networks
Today, NLP is quickly transforming into a powerful and robust tool that allows for a deeper understanding of the context on mobile. Joseph shared a demo of how their predictive models use ELMo, an NLP language processing algorithm, which is able to recognize an apps’ context (its features, mechanics, and themes) and create a vector representation of the app store description, which encompasses how users would classify the app. In short, this allows Persona.ly to collect the data from the app store descriptions and measure the contextual distance between apps and games. This helps find suitable apps to advertise the promoted game on and increase installs and ROAS.
Achieving quality and scale with iOS14.5
As we transition into iOS14.5, it’s essential to have enough scale to understand how to measure value. With Apple’s new framework accompanied with many limitations. Even getting the conversion value to understand performance has become challenging, and it will take some time to adapt to this new playbook. Adding to the complexity of these tasks, the billing cycle and understanding your ROI will also take a couple of months to gauge.
“It will take time for advertisers that spend at scale to understand the impact, as well as keeping some money aside for a rainy day.”
– Joseph Iris, Persona.ly
Joseph’s look ahead:
There’s going to be many iterations and more extensive research around optimizing the conversion value to actually represent value. It’s not as intuitive as one might think because of the limitations and the “security by obscurity” approach that are put in place by Apple to disconnect the two. The balance in prices and this loss of accuracy will align. The hunger is still there, growth is still the mindset – it’s just a playing field that is much more abstract.
With iOS14.5 already here, Joseph breaks down the sustainable long-term and alternative paths that Persona.ly are taking to adapt to the changes and work closely with advertisers to adapt and drive success.
Join us for the hands-on discussion we had on Digital Fyber Focus with more on:
- How app store descriptions make a contextual difference
- NLP machine learning techniques to represent users and apps
- Early insights on LAT inventory and user acquisition
- Rethinking attribution windows, spend and the eCPM shift
- Adjusting advertisers expectations on spend